From owner-freebsd-current  Thu Aug 22 00:58:49 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id AAA25582
          for current-outgoing; Thu, 22 Aug 1996 00:58:49 -0700 (PDT)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.7.5/8.7.3) with SMTP id AAA25577
          for <current@freebsd.org>; Thu, 22 Aug 1996 00:58:46 -0700 (PDT)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.12/8.6.9) id RAA20229; Thu, 22 Aug 1996 17:53:37 +1000
Date: Thu, 22 Aug 1996 17:53:37 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199608220753.RAA20229@godzilla.zeta.org.au>
To: current@freebsd.org, erich@lodgenet.com
Subject: Re: does HAVE_FPU really work?
Sender: owner-current@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

>(ttyp1@jake)# pwd
>/usr/src/lib/msun
>(ttyp1@jake)# make
>cc -O -D_MULTI_LIBM -D_POSIX_MODE -D_IEEE_LIBM -c /usr/src/lib/msun/src/e_acos.
>c -o e_acos.o

>shouldn't it be compiling /usr/src/lib/msun/i387/e_acos.S ???

Yes.

>Oh, well, make does a lot of stuff under the hood, so I proceeded to let
>it compile, but the resultant library libm.so.2.0, has exactly the
>same md5 as without HAVE_FPU defined --???

>then trying `make -ds e_acos.o', I get amongst other stuff:

>SuffFindDeps (e_acos.o)
>        trying e_acos.s...not there
>        trying e_acos.S...got it
>        using existing source /usr/src/lib/msun/src/e_acos.c
>        applying .c -> .o to "e_acos.o"

>looks like it finds e_acos.S, but uses e_acos.c anyway.
               didn't find

It shouldn't use e_acos.c, since e_acos.c isn't among the source
files if HAVE_FPU is defined (see the Makefile).

Bruce